Universal-2-TF: Robust All-Neural Text Formatting for ASR

TL;DR

This paper presents a neural text formatting model for ASR that improves accuracy, efficiency, and robustness across various languages and domains, outperforming traditional methods.

Contribution

It introduces a novel two-stage neural architecture for all-neural text formatting in ASR, reducing hallucinations and computational costs.

Findings

Superior TF accuracy demonstrated in evaluations

Enhanced computational efficiency over existing methods

Improved perceptual quality in ASR outputs

Abstract

This paper introduces an all-neural text formatting (TF) model designed for commercial automatic speech recognition (ASR) systems, encompassing punctuation restoration (PR), truecasing, and inverse text normalization (ITN). Unlike traditional rule-based or hybrid approaches, this method leverages a two-stage neural architecture comprising a multi-objective token classifier and a sequence-to-sequence (seq2seq) model. This design minimizes computational costs and reduces hallucinations while ensuring flexibility and robustness across diverse linguistic entities and text domains. Developed as part of the Universal-2 ASR system, the proposed method demonstrates superior performance in TF accuracy, computational efficiency, and perceptual quality, as validated through comprehensive evaluations using both objective and subjective methods. This work underscores the importance of holistic TF models in enhancing ASR usability in practical settings.